Development And Validation Of A Nomogram To Predict Anastomotic Leakage After Esophagectomy For Cancer

Objectives: This study aimed to identify variables associated with anastomotic leakage after esophagectomy and established a tool for anastomotic leakage predictionNomograph,which can assess the risk of anastomotic leakage after esophagectomy,and to provide help for clinical decision-making.Methods: The clinical data of patients with esophageal cancer who underwent radical esophagectomy from January 2018 to June 2020 in the Affiliated Hospital of Qingdao University were collected retrospectively.Esophagectomy Complications Consensus Group(ECCG)standard were used to define and grade anastomotic leakage;25preoperative and intraoperative parameters including gender,age,ASA classification,comorbidities,operation time,anastomotic position,and anastomotic technique were included for analysis.SPSS Version 23.0 and Empower Stats software were used for establishing a nomogram after screening relevant variables by univariate and multivariate Logistic regression analyses.The established nomogram was evaluated by the Receiver Operating Characteristic curves(ROC),calibration curve and decision curve analysis,which was verified by 1000 bootstrap resamples method.Results: A total of 604 eligible patients with esophageal cancer were collected,including 564 males(93.38%)and 40 females(6.62%),with an average age of 63.5±8.1years.The number of patients with tumors located in the lower thorax was the largest,accounting for about 57.78%,squamous cell carcinoma is still the main histological type(479,79.3%),and 95.7%(578)of patients achieved R0 resection.There were 51 patients with anastomotic leakage,accounting for 8.4% of the total,including 5(0.8%)Type Ⅰ,41(6.8%)Type Ⅱ and 4(0.7%)Type Ⅲ cases.The average length of hospital stay of patients with anastomotic leakage was 50(34-76)days.The average length of hospital stay of patients without anastomotic leakage was 22(17-28)days,the difference was statistically significant(P < 0.001).Univariate Logistic regression analysis found 9 risk factors that may be related to anastomotic leakage,including smoking index,ASA,neoadjuvant therapy,tumor location,operation time,surgical method,anastomosis method,anastomosis location and prognostic nutritional index.Incorporating the above parameters into the multivariate logistic regression analysis found that smoking index,anastomotic position,anastomotic method,prognostic nutritional index and ASA were independent risk factors for anastomotic leakage.The area under the curve(AUC)is 0.764(95%CI: 0.696-0.832),which shows that the model has a good degree of discrimination,indicating that the established nomogram has a good discriminative ability in the prediction of anastomotic leakage.The calibration curve shows a good fit between the predicted probability and the actual observed probability.During the internal verification,a high consistency index of 0.767(95%confidence interval,0.699-0.835)was maintained,which supports that the nomograph could be widely used in the prediction of anastomotic leakage.In addition,the decision curve analysis shows that compared with the original treatment strategy,the use of this model to identify patients with a higher incidence of anastomotic leakage and then take intervention measures may gain more clinical benefits.Conclusions: An anastomotic leakage risk prediction model based on individualized scores based on the patient’s smoking index,ASA score,anastomotic position,anastomotic method,and prognostic nutritional index can be used to screen out highrisk patients who may develop anastomotic leakage,which is useful to clinicians focus on screening and follow-up of patients,and have certain reference value for the formulation of preoperative and postoperative intervention strategies.